Evolutionary dynamics of neoantigens in growing tumours

Abstract

Cancer evolution is driven by the acquisition of somatic mutations that provide cells with a beneficial phenotype in a changing microenvironment. However, mutations that give rise to neoantigens, novel cancer-specific peptides that elicit an immune response, are likely to be disadvantageous. Here we show how the clonal structure and immunogenotype of growing tumours is shaped by negative selection in response to neoantigenic mutations. We construct a mathematical model of neoantigen evolution in a growing tumour, and verify the model using genomic sequencing data. The model predicts that, in the absence of active immune escape mechanisms, tumours either evolve clonal neoantigens (antigen-'hot'), or have no clonally-expanded neoantigens at all (antigen-'cold'), whereas antigen-'warm' tumours (with high frequency subclonal neoantigens) form only following the evolution of immune evasion. Counterintuitively, strong negative selection for neoantigens during tumour formation leads to an increased number of antigen-warm or -hot tumours, as a consequence of selective pressure for immune escape. Further, we show that the clone size distribution under negative selection is effectively-neutral, and moreover, that stronger negative selection paradoxically leads to more neutral-like dynamics. Analysis of antigen clone sizes and immune escape in colorectal cancer exome sequencing data confirms these results.Overall, we provide and verify a mathematical framework to understand the evolutionary dynamics and clonality of neoantigens in human cancers that may inform patientspecific immunotherapy decision-making.

Tumor neoantigen heterogeneity thresholds provide a time window for combination immunotherapy

“…However, only recently have the models included genetic instability and neoantigen load as key drivers of immune recognition [22]. Recent approaches are also including neoantigen evolution [16] or subclonal structure and immunodominance effects [17] as parameters in the tumor-immune interaction. Can we make use of these results to understand the role of neoantigen clonality in immune control of malignancy?…”

“…How does immune attack and cancer death correlate with neoantigen composition? From the set of neoantigens presented by a clone, only a very few are of significant immunogeneicity and elicit a strong T cell response [17], thus cancer cell death is expected to correlate with the most immunogenic antigen in their repertoire.…”

“…Previous research has already studied the neoantigenic capacity of epitopes, termed neoantigen fitness [17]. However, further work indicates that even highly recognizable antigens fail at inducing T cell response if they are not present in a sufficient fraction of the tumor [14].…”

“…Prior modeling exercises have succesfully evaluated tumor evolution and response to immunotherapy under this perspective [17], but without explicitly considering heterogeneity in the antigen landscape. Furthermore, the process of evaluating the most dominant antigen in each clone, its recognition potential, and its distribution among the rest of subclones relies on bulk data processing where sampling biases due to localized biopsies [23] and imperfect antigen prediction [24] introduce a high level of noise.…”

“…Recent mathematical insight has emerged from the study of neutrality of neoantigen evolution as well as the effect of selective pressures towards immune escape [16]. Further research has proven it possible to characterize immunotherapy prognosis by quantitatively assessing the immunogeneicity of subclonal epitope composition [17]. However, how the degree neoantigen heterogeneity correlates with the immune response remains to be studied from a mathematical perspective.…”